Valuing bundled attributes: a latent characteristics approach.

摘要

A method believed to address both the issues of multicolinearity in bundled attributes and the need to represent bundled attributes as "characteristics" that directly enter the consumer's utility function, is described. Data on visitor's origin andtrail attributes were collected from 4778 visitors to 46 trails in 20 wilderness areas of the Southeastern United States (eastern Tennessee, northern Georgia, and in the Smoky Mountains of North Carolina). A model of characteristics was estimated in which there are 2 directly observed trail characteristics (elevation change and isolation) and 2 latent bundled characteristics (upland habitat and riparian habitat). The results of the latent analysis showed that upland and riparian appear to be latent bundles of characteristics, however, some of the indicator variables are not highly significant explanatory variables for latent characteristics.